gergely/astrognome
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Added chat log on the topic of 2D drawing

Browse Source
This commit is contained in:
Gergely Polonkai
2013-09-03 18:12:02 +02:00
parent 5cfb27d3cd
commit e082c0fe07
1 changed files with 52 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

52
graphics-help-from-irc.txt Normal file
View File

@@ -0,0 +1,52 @@
<W00d5t0ck> hello
<W00d5t0ck> what is the best way to create still 2D graphics?
<desrt> cairo
<desrt> it has a fairly nice interface like "go to this point, then to this point, draw a line, draw an arc" and so on
<W00d5t0ck> can it handle external images?
<desrt> like pngs, etc?
<W00d5t0ck> I have svg, as it scales better
<W00d5t0ck> but basicly, yes
<desrt> you can use librsvg to render svgs to cairo, in vector form
<desrt> if you go through a different image loader library (like GdkPixbuf) then you'll end up rasterising it first
<W00d5t0ck> the first version is more to my likes
<desrt> you might find this helpful: http://cgit.freedesktop.org/~cworth/svg2pdf/tree/svg2pdf.c
<desrt> shows how to render an svg to a cairo pdf surface
<desrt> but you could just as well use it for other cairo surfaces, including the ones you get out of gtk
* nkoep kilépett (Read error: 145 (Connection timed out))
<W00d5t0ck> thank you
<desrt> are you making something that you want to end up on the screen eventually?
<W00d5t0ck> yes
<W00d5t0ck> and possibly to save it to an image file later
<desrt> depending on your app, you'll either want to create a cairo image surface and draw to it, then draw the image surface to the screen from your gtk draw() function
<desrt> or just draw directly to gtk's cairo context from your draw function
<desrt> cairo_surface_write_to_png() will help with that
* yoseforb kilépett (Remote closed the connection)
<W00d5t0ck> I'll take a look then
* yoseforb (~yoseforb@164.138.127.79) csatlakozott ide: #gtk+
<W00d5t0ck> and if I want to get a bit farther? Like I want to add some actions to the svg parts? E.g. tooltips should appear when I hover a specific "sub-image"
<desrt> that would be complicated
<W00d5t0ck> I was more than sure :)
<desrt> particularly if you wanted to do it for named parts of the svg file...
<desrt> if you want to do it for a particular hard-coded coordinate you could use a motion notify event and a handler that does a popup when the mouse is over the area you care about
<desrt> but i don't think rsvg has any mechanism for saying "the box with xml id 'foo' is at these coordinates"
<W00d5t0ck> no, the svg files I load are small icons actually
<W00d5t0ck> no named parts
<desrt> probably would not be too hard then
<W00d5t0ck> they are svg onlf for scalability
<desrt> from the Gtk side, you're probably going to want to use a GtkDrawingArea
<W00d5t0ck> I've already figured out that, I'm currently digging my nose into Gtk documentation to see how to draw on it
<desrt> then do gtk_widget_add_events() on it for GDK_ENTER_NOTIFY_MASK | GDK_LEAVE_NOTIFY_MASK | GDK_POINTER_MOTION_MASK
<chpe> actually there is an API for that in rsvg, but it's horribly slow
<desrt> W00d5t0ck: connect to the "draw" signal on the drawing area
<desrt> it will hand you a cairo_t as one of the arguments
<desrt> draw on it....
<chpe> well, id -> rect, but not the other way
<desrt> chpe: neat. that's very useful.
<desrt> W00d5t0ck: what language are you using?
<desrt> if it's anything other than C, you almost certainly want to subclass GtkDrawingArea
<W00d5t0ck> I'm only familiar enough with C
<desrt> and if it is C, and you know how to subclass, i'd still recommend doing it that way, and overriding the 'draw' function in the subclass
<W00d5t0ck> I've done subclassing back in Gtk 2, with C
<desrt> it's the same deal still
<W00d5t0ck> so I think it won't be a problem
<desrt> 'cept gtk2 didn't have 'draw'